On Thu, 2004-01-15 at 01:57, Poppe, Troy wrote: > The linux image will not respond to telnet, ssh, or ping. The CPU is being spin > so quickly that nothing is getting handled other than those java processes (and I > wonder if any real processing is being done with them). > > Technically speaking, the jvm isn't crashing... Currently, the linux kernel > imposes no limits on the user space. (This is a test instance...) > > My curiosity lies exactly where you suggest, a cpu loop. The question to me is > where this cpu loop is. I'm not aware of anything in JBoss that could/would > cause a tight loop like I am seeing. (I've been running JBoss on Linux in x86 > and Windows for almost a year and a half now and never seen anything quite like > this.) I wonder if the starvation of physical memory is having an impact on the > jvm that is undesirable (your case of a system-wide gc with a heavily swapped jvm > poses an interesting question of how this would be handled). As of right now, I > have no data to direct me other than to keep trying to get this thing to crash. > > I will note that I am currently running the IBM Java2 1.3.1 R6 for s390, and the > instance has been stable for over 12 hours (idle). It may very well be a problem > in the 1.4.1 version from IBM. > > Again, Adrian, thanks for your continued help. Do you know of any one who is > running JBoss on Linux for z/Series that might be able to shed some light on > this? >
No I don't. I know Holger Baxmann was playing with JBoss on the z/Series about a year ago. He hovers around these lists sometimes. Regards, Adrian > Thanks. > > Troy Poppe > > ------ > > Subject: Re: [JBoss-user] JBoss 3.2.3 problems running in Linux for z/Seri > es > From: Adrian Brock <[EMAIL PROTECTED]> > To: [EMAIL PROTECTED] > Organization: > Date: 14 Jan 2004 22:10:28 +0000 > Reply-To: [EMAIL PROTECTED] > > So can you ping or telnet into the linux image? That will tell you > whether linux is really dead. > > In principle a jvm should never crash, if it does it is a bug in the > jvm. > Of course the OS may decide to kill the process by sending it a signal > (e.g. exceeded the cpu limit of ulimit or a segfault) but that shouldn't > bring down the OS. The jvm should report the signal with a dump. > > 100% cpu utilization implies a cpu loop somewhere. > Your top output showed this was mostly system cpu, > i.e. inside the OS. It could be the jvm doing a system call in a tight > loop. > > Have you checked /var/log/messages to see whether Linux reported > a problem? > > Regards, > Adrian > > > > ------------------------------------------------------- > This SF.net email is sponsored by: Perforce Software. > Perforce is the Fast Software Configuration Management System offering > advanced branching capabilities and atomic changes on 50+ platforms. > Free Eval! http://www.perforce.com/perforce/loadprog.html > _______________________________________________ > JBoss-user mailing list > [EMAIL PROTECTED] > https://lists.sourceforge.net/lists/listinfo/jboss-user -- xxxxxxxxxxxxxxxxxxxxxxxx Adrian Brock Director of Support Back Office JBoss Group, LLC xxxxxxxxxxxxxxxxxxxxxxxx ------------------------------------------------------- This SF.net email is sponsored by: Perforce Software. Perforce is the Fast Software Configuration Management System offering advanced branching capabilities and atomic changes on 50+ platforms. Free Eval! http://www.perforce.com/perforce/loadprog.html _______________________________________________ JBoss-user mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/jboss-user
